- 1 I assume that the conceptual analysis in legal positivism consists in a rational reconstruction of (...)
1The purpose of this article is to provide a critical overview of select solutions to a certain problem that confronts legal positivism, namely, the problem of moral evaluations in legal interpretation. The central question concerns whether the involvement of moral evaluations in legal interpretation can be reconciled with the positivistic separation thesis.1
- 2 This general formulation has been criticized inter alia by Gardner 2001, Green and Adams 2019, and (...)
2In its original general formulation, the separation thesis claims that any connections between law and morality are contingent and not conceptually necessary (Hart 2016).2 The separation thesis is a theoretical claim that says something about the law, not about legal theory. It must be distinguished from the meta-theoretical claim that legal theory is morally neutral and has no justificatory aims (Hart 2012: 240). The latter claim is referred to as “the neutrality thesis” and constitutes the central claim of so-called methodological positivism.
- 3 For a detailed analysis of possible understandings of the separation thesis, see Pino 2014 and Mora (...)
3The separation thesis comes in several versions.3 I am interested in its weak version, which Marmor formulates as follows: “[D]etermining what the law is does not, necessarily or conceptually, depend on moral or other evaluative considerations about what it ought to be in the relevant circumstances” (Marmor 1999: 135). This version is considered weak because it does not exclude that determining what the law is may contingently depend on moral considerations. The strong version of the separation thesis claims that determining what the law is never depends on moral considerations. I have chosen to address the weak version because if it turns out that the necessary involvement of moral considerations in legal interpretation falsifies the weak version, then a minori ad maius, all of its stronger versions are also false.
4The separation thesis is negative in that it makes a claim about what the law does not necessarily depend on, but it says nothing about the positive aspect, namely how the law is determined. This positive aspect is supplemented by the positivistic social thesis, which also appears in various versions. Let us consider just two:
[T]he existence of law requires some form of social practice that includes judges and ordinary citizens, and this social practice determines, in any given legal system, the ultimate sources of law, or the ultimate criteria or ultimate tests for legal validity (Hart 2016: 461).
Legal facts are ultimately determined by social facts alone (with exclusion of any moral facts) (Shapiro 2011: 27).
5To specify, legal facts are facts about legal content (“x is permitted”, “y is forbidden”, “A has the right to do a” etc.).
- 4 See also Moreso 2001: 38.
- 5 On the foundational theses of legal positivism, see Himma 2019: 66-76.
- 6 Gardner 2001: 4.
6The relationship between those versions of the social thesis and the separation thesis remains undermined. Shapiro’s social thesis entails the separation thesis in its strong version. If the law is ultimately determined by social facts alone, then morality plays no role in its determination. The matter is not so clear, however, with respect to the Hartian thesis. In his list of fundamental claims of legal positivism, the separation thesis and social thesis are separate items (Hart 2016: 461),4 but this does not mean that they are unrelated. The rule of recognition is undoubtedly a social rule and, as such, is solely a matter of social practice and not a matter of morality. Therefore, accepting the Hartian social thesis commits us to the separation thesis, at least in its weak version.5 We must stress that the separation thesis, as a consequence of the social thesis, excludes not only moral considerations as necessary determinants (or co-determinants) of the law but also any other nonmoral evaluative considerations (such as religious, aesthetic, economic, etc.).6
- 7 Another question regards what sort of metaphysical determination is meant here.
7The verb “to determine” as used in the social thesis and the separation thesis gives rise to a certain ambiguity. Both theses can be interpreted as either metaphysical or epistemic claims. Metaphysically, both theses claim is that the law is ultimately metaphysically determined solely by social facts without any moral facts.7 Epistemically, both theses answer the question of how the law is cognized. Thus, the social thesis claims that to identify the law, we must consider social facts. The separation thesis (in its weak version) claims that no moral considerations are necessary for this purpose, although they may be contingently applicable. Below, I adopt the epistemic interpretation and use the verb “to determine” in the epistemic sense.
- 8 In continental legal theory, an important distinction is made between norm-formulation and norm. A (...)
8A comparison of the two versions of the social thesis referenced earlier shows that we need to distinguish three levels: (i) the level of social facts or social practice, (ii) the level of valid legal norms, and (iii) the level of legal content.8 Hart’s thesis refers to the relationship between the first and the second levels (social practice determines which norms are legally valid). In contrast, Shapiro’s thesis refers to the relationship between the first and the third levels (social facts determine legal content). Obviously, the content derived from valid legal norms is meant here.
- 9 I skip here the problem of material validity (coherence with the relevant higher-rank norms), which (...)
9The separation thesis seems unproblematic within the positivistic camp with respect to the move from the first to the second level. Positivists claim that each legal system is based on a social rule of recognition, the function of which is to determine which social facts are sources of law and, in this manner, to determine the criteria of legal validity in the system. Certainly, the criteria set out by the rule of recognition may be somewhat vague, or there may arise a controversy concerning whether those criteria have actually been satisfied. This particularly applies to inclusive positivism, which claims that such criteria may be moral (or more generally, content-based), provided that in the contingent social practice constituting the conventional rule of recognition, such criteria are actually applied by judges. If that is the case, the application of those moral criteria (but not their general applicability) may be disputed.9
- 10 Spaak (2021: 446) correctly stresses the distinction between two levels of legal thinking: one ques (...)
10The problem with the separation thesis centres on the move from the level of valid legal norms to the level of legal content. Knowing which norms are legally valid does not provide immediate access to legal content. Answering questions relating to legal content requires interpretation.10 It is frequently claimed that legal interpretation involves evaluative considerations, including moral evaluations, and that we are unable to answer interpretative questions without explicit or implicit recourse to values. If this claim is true, it jeopardizes the separation thesis. If the identification of legal content necessarily involves moral evaluations, then legal content is never fully determined by social facts alone, and therefore the truth (or acceptability) of propositions about legal content depends on the truth (or acceptability) of moral propositions. Thus, the separation thesis (even in its weak version), as applied to the move from the second to the third level (from valid legal norms to legal content), appears to be false.
11The falsity of the separation thesis (if it is actually false) would have a powerfully adverse consequence for legal positivism. If the social thesis implies the separation thesis, the falsity of the latter triggers the falsity of the former (modus tollendo tollens). Some positivists claim that the social thesis is the only distinctive proposition of legal positivism. Therefore, if the falsity of the social thesis were demonstrated, legal positivism would be defeated.
12Yet, we must note that such an argument against legal positivism is based on certain assumptions. The first assumption is metatheoretical, in that it claims that a theory of interpretation is a necessary part of legal positivism qua a theory of law. If it is not such a part, then even the necessary involvement of evaluations in legal interpretation is coherent with the separation thesis, since positivism is solely a theory of validity. The second assumption is that the law, as a matter of necessity, requires interpretation — in other words, it is impossible to answer questions relating to legal content without interpretation. If this assumption is false, the problem disappears because in those cases where no interpretation is required, legal content may be identified based on social facts alone, which is sufficient for defending the weak version of the separation thesis. The third assumption is that legal interpretation necessarily encompasses moral evaluations. But if such an evaluation is required only contingently, the argument fails because the separation thesis, in its weak form, claims only that the law does not necessarily depend on moral evaluations and does not deny that it may contingently depend on them.
13Legal positivists have the following options. First, they may reject the first assumption and assert that legal positivism is solely a theory of legal validity (and does not include a theory of legal interpretation) and, therefore, any involvement of moral evaluations in legal interpretation is irrelevant from the perspective of the social thesis and the separation thesis. Second, they may reject the second assumption and argue that it is at least sometimes possible to identify legal content without the need for interpretation. Third, they may reject the third assumption and argue that interpretation does not involve (or only contingently involves) moral evaluations.
- 11 One important caveat: I will not discuss the stance of Joseph Raz. Due to the complexity of this st (...)
- 12 Obviously, the main opponent of positivism with respect to the matters discussed in this paper is t (...)
14In the following, I briefly discuss the merits and demerits of these three options. I refer to certain selected accounts proposed in the legal-theoretical literature. The overview presented below is not complete, but I do believe that the selected accounts are representative of the main lines of thought within legal positivism.11 The first two options are discussed by using one theoretical account of each of them as examples. The third option is discussed by reference to two accounts, which seem to me fundamentally different.12
- 13 Chapter VII of the Concept of Law deals almost exclusively with the open texture in law. Obviously, (...)
15Positivists frequently contend or tacitly assume that legal positivism is solely a theory of legal validity, and therefore, both the social thesis and the separation thesis apply exclusively to the identification of valid legal norms but not to the determination of legal content. The separation thesis claims that the validity of law does not necessarily depend on its conformity with morality. Hart’s position on this matter is unclear. Although Hart did not explicitly discuss the problem of interpretation in the context of the separation thesis, he was of the view that the matters of interpretation and validity are separate. The doctrine of the rule of recognition is a theory of validity, since the rule of recognition determines the ultimate criteria of validity. Hart was fully aware that the interpretation of law is value-laden (Hart 2012: 204), but he did not say much about interpretation.13 He did, however, assert that
Even if the judges [...] did in fact settle questions of meaning in […] [a] partly evaluative way, this would be something for the general descriptive theorist to record as a fact [...] Description may still be description, even when what is described is an evaluation (Hart 2012: 244).
16On the other hand, however, in the Postcript Hart says:
according to my theory, the existence and content of the law can be identified by reference to the social sources of the law […] without reference to morality except where the law thus identified has itself incorporated moral criteria for identification of the law (Hart 2012: 269).
- 14 Inclusive positivist will, however, argue that ultimately the validity depends on social facts, sin (...)
- 15 In a similar spirit, Burazin and Ratti argued that the rule of recognition does not necessarily enc (...)
17Torben Spaak has recently advocated for, and elaborated in detail on, the view that legal positivism is solely a theory of validity (Spaak 2021). He takes as his starting point the distinction between two levels of legal thinking: “It is one question whether a legal norm [...] is legally valid, that is, exists as a legal norm, and another question just how this legally valid norm is to be interpreted and applied” (Spaak 2021: 446). He then argues that this distinction is natural, since it is the distinction “between the determination of the object of interpretation and determination of the (correct) interpretation of the object of interpretation” (Spaak 2021: 447). The social thesis — and therefore the separation thesis as its consequence — applies exclusively to the first level, which encompasses the move from social facts to valid legal norms. Exclusive positivists say that the validity of law does not depend on its conformity with morality, but is solely based on social facts constituting the sources of law. Inclusive positivists say that the validity may contingently, but does not necessarily have to, depend on morality. But they also claim that validity always ultimately depends on social facts, since whether morality is a criterion of validity depends on the rule of recognition, which is a social rule.14 On the other hand, interpretative questions such as which canon of interpretation should prevail, are, pursuant to Spaak, not factual but evaluative (typically moral) questions. A judge answering such a question must typically use moral criteria. Therefore, the determination of legal content is not solely based on the cognition of social facts.15
- 16 I owe this reservation to one of the anonymous reviewers.
18If positivism is solely a theory of legal validity and has nothing (or little) to say about interpretation, it does not seem to be especially interesting since it is unable to provide a sufficient theoretical basis for an explanation of legal practice, which seems relevant for the epistemic reading of the social thesis.16 But, Spaak refutes this objection, arguing that once we have solved the question of validity, we have to deal with a set of uninterpreted valid norms. This set is ultimately determined solely on the basis of social facts because the rule of recognition that sets out the criteria of validity is a social rule. Spaak argues, “[U]ninterpreted norms [...] can and do guide human behaviour for the simple reason that law-appliers in general rarely deviate from the plain meaning of the norms” (Spaak 2021: 462). Therefore, from the perspective of those whose behaviour is subject to legal regulation, the identification of valid “uninterpreted” norms is, in most cases, sufficient for practical purposes. The addressees of legal regulations are rarely faced with interpretative problems. Rather, interpretation is a matter of judicial practice in hard cases, and the number of such cases is limited (Spaak 2021: 453).
19For purposes of the above argument, Spaak must assume that ascribing the plain meaning to the term “uninterpreted norms” is not an interpretative choice based on evaluative criteria. This assumption is disputable, and I discuss it further below. Here, however, note that Spaak’s thesis that positivism is solely a theory of validity does not require the adoption of this assumption. Therefore, one may agree with Spaak’s thesis concerning the scope of legal positivism and disagree with his claim that certain norms do not require interpretation.
20For Spaak’s conception, positivism is a sort of “minimum theory”, as it explains only “obvious law”. As a theory of legal validity, positivism qua legal theory has nothing to say about legal interpretation. There can be no distinctively positivistic theory of interpretation.
- 17 This section is an extended and substantially modified version of my paper (Gizbert-Studnicki 2024) (...)
- 18 Since space constraints prevent a deeper discussion of this philosophical problem, I only discuss c (...)
21Some legal positivists reject the second assumption and argue that the determination of legal content only sometimes requires interpretation.17 The plausibility of the second assumption is a matter of general philosophical controversy relating to the problem of the role of interpretation in textual and verbal comprehension. The stance defined as “interpretative universalism” claims that all comprehension is necessarily a matter of interpretation. Interpretative universalism is accused of neglecting some obvious features of ordinary understanding (Patterson 2006). The famous Wittgensteinian problem of rule-following is closely linked to this issue.18
- 19 On philosophical problems related to distinguishing understanding and interpretation, see Sandro 20 (...)
22Marmor and other adherents of the view that the identification of legal content does not necessarily involve interpretation agree that each interpretation is value laden. Marmor argues, however, that interpretation is only an exception to the ordinary understanding of a legal text (Marmor 2005a).19 We must engage in interpretation only if doubt arises regarding how to apply a rule. If a legal text is clear, no interpretation is needed. Marmor agrees that interpretation always requires a choice based on evaluative criteria; in most cases, however, legal texts are linguistically clear and allow legal officers and ordinary citizens to identify legal content without making any such choice. One’s understanding of a norm is manifested in the ability to apply it in standard cases. Understanding a norm does not need to involve a statement of the necessary conditions for its every possible application. A norm may be understood with respect to one certain case and remain unclear in another less certain one.
23The adherents of this view usually recognize that the doubts that trigger the need for interpretation may be caused not only by linguistic defects (such as ambiguity or vagueness) but also by extralinguistic factors (such as obvious immorality or absurd consequences from applying a norm according to its literal meaning in a given case). Nonetheless, they believe that the need for interpretation is an exception rather than a rule. It is worth stressing that the precondition for such extralinguistic doubts is a prior understanding of the rule in question. We cannot say that a rule is immoral or absurd if we do not understand its linguistic meaning.
24Marmor argues that
[the] separation thesis necessarily involves the assumption that judges can (at least in some standard cases…) identify the law and apply it without reference to considerations about what the law ought to be in the circumstances [...]. The distinction between the law as it is and the law as it ought to be entails a parallel distinction between the activity of simply understanding the law and just applying it and modifying or creating it (Marmor 2005a: 95).
25He further claims that interpretation is always a (partly) creative activity because it always “adds something new, previously unrecognized to that which is being interpreted” (Marmor 2005a: 125). Therefore, it aims at establishing what the law ought to be, and “legal positivism cannot accept the view that law is always subject to interpretation” (Marmor 2005a: 125). On this basis, Marmor claims that positivism is committed to a distinction between easy cases, where the law is simply understood, and hard cases, where interpretation of the law is needed; otherwise, the conceptual separation between law as it is and law as it ought to would be untenable. This distinction is fundamental for legal positivism and must have a basis independent from the legal positivist doctrine (otherwise, we would be question-begging) (Marmor 2025a: 95-6). This is an important point: The positivistic separation thesis does not per se constitute an argument for distinguishing understanding and interpretation. If Marmor is right, the separation thesis must assume this distinction, which should be separately justified.
- 20 On the different senses of the phrase “linguistic understanding”, see Grodniewicz 2021.
26It seems that Marmor conceives of understanding as a mental state resulting from a cognitive process governed by the linguistic competence of the recipient of an utterance.20 In this sense, understanding is not an action but is something that happens to us. In contrast, interpretation is an action — or a chain of actions — that is at least partially creative, insofar as it necessarily involves a choice. Without a choice, the doubt triggering the need for interpretation cannot be eliminated. Such a choice is based on normative criteria (moral, religious, economic, or aesthetic). Another difference between interpretation and understanding is that when interpreting a norm, we paraphrase its original wording. Understanding a norm does not involve such paraphrasing but instead takes the wording of a norm as it is.
- 21 Tobia, Slocum, and Nourse (2023: 19) write “Textualism claims to promote rule of law values (e.g., (...)
- 22 A version of this argument is Fuller’s account in his criticism of Hart. Fuller argues that in the (...)
27Against this view, some occasionally argue against the idea that understanding based solely on linguistic competence excludes implicit evaluative choice. The arguments are based on the idea that the very selection of the linguistic meaning as the basis for determining legal content is grounded in moral evaluations — in particular, the values of legal transparency and the protection of expectations. Such a claim is characteristic of textualists,21 as they argue that no method of interpretation can be value neutral. Each method of interpretation involves values (even if such values are only tacitly assumed). Even one’s simple understanding based only on linguistic competence involves an implicit normative choice.22
- 23 On the concept of the ordinary meaning of rules, see Slocum 2014.
- 24 In the decision District of Columbia v. Heller, the US Supreme Court said: “In interpreting this te (...)
28I do not believe that such argumentation against discriminating between understanding and interpretation is sound. First, outside legal contexts, the literal linguistic meaning is usually ascribed to written and spoken utterances in everyday communication by default.23 In standard cases, it is simply a matter of instrumental rationality for a speaker to use words according to their standard linguistic meanings. Similarly, the legislature tends to speak to addressees of law in the language they are familiar with. Therefore, the choice to ascribe the literal meaning to a legal text by default is not based on any moral choice. Still, it is a simple consequence of instrumental rationality, which in my view is not a moral principle.24 Here, by “literal meaning”, I mean the meaning determined by the conventions of language (Slocum 2014: 302).
29Second, we can invoke a well-known argument. The view that the understanding of a linguistic expression (word, clause, sentence, etc.) always requires interpretation, triggers a regressus ad infinitum because the result of interpretation must be formulated through the use of another linguistic expression of the form “L means X”. If the outcome of interpretation is again a linguistic expression, it also must be interpreted, and so on. Therefore, the process of interpretation can never be completed. To avoid such a regressus, we must stop somewhere and say, “This expression is understandable without any further interpretation”. But this means that understanding without interpretation must be conceptually possible.
30To summarize Marmor’s position, he claims that if the case under consideration constitutes a standard example that unquestionably falls within the clear literal meaning of the norm, no interpretation occurs. This is sufficient to defend the weak version of the separation: It is neither conceptually nor empirically necessary to engage in moral evaluations in the identification of legal content. This is not a normative, but a conceptual claim. Marmor does not contend that judges are compelled to decide in accordance with the literal meaning of the norm to be applied. The question of whether and to what extent judges are bound by the literal meaning of the norms they apply is a normative question. Therefore, legal positivism, as a descriptive theory, has no ambition and no resources to answer this question. It is only a statement of fact that judges, in most cases, determine legal content based on the literal meaning of valid legal norms without engaging in interpretative activity. The need for interpretation arises when the language of law is ambiguous, vague, or indeterminate in any other respect. In addition, judges sometimes refuse to apply a rule in accordance with its literal meaning and instead modify this meaning if the adoption of the literal meaning would lead to consequences that they consider morally wrong or absurd. But such a departure from the literal meaning assumes that the norm must have been previously understood. One cannot say “The norm R is morally wrong” if one does not understand R. Therefore, departing from the literal meaning in the process of interpretation by giving priority to extralinguistic canons also requires a pre-interpretative understanding of the norm in question.
31Chiassoni, in his critique of Marmor’s account, claims that it is based on the premise that the law, as it is for any legal order, is tantamount to the clear, ordinary meaning of the norms valid in such an order (Chiassoni 2008: 260). If a norm does not have such a clear meaning, the identification of its legal content requires interpretation, which always involves evaluation. Chiassoni argues that the adoption of such a premise is based on a normative choice, and this premise simply gives absolute priority to the literal canon of interpretation over all other canons (Chiassoni 2008: 260).
32That is certainly not Marmor’s view. Marmor, in his critique of textualism, explicitly rejects the premise ascribed to him by Chiassoni (Marmor 2005b). Leaving aside the normative arguments Marmor deploys, the premise ascribed to him makes a normative claim, whereas Marmor’s account is descriptive. Further, the understanding reflecting the literal meaning is usually unreflective and automatic, and is therefore not based on any canon, since the application of a canon must always be reflective. It is true that the literal canon has some moral connotations (its application is linked to the values of legal certainty, etc.), but the need to apply this canon (and consequently prioritize the literal meaning over alternatives identified on the basis of other canons) is triggered only if a doubt arises. Ascribing the literal meaning in the absence of doubt is the outcome of an unreflective and automatic understanding and, as such, is not based on any moral choice.
33Marmor may be right insofar as his claim is purely descriptive. It is plausible that in real-life situations, people frequently obtain access to legal content in an unreflective manner. It is plausible that in many cases, judges just rely on the wording of the norm to be applied without undertaking any interpretative effort. If the law says that the deadline for lodging a motion is 14 days, there is usually, in standard contexts (but not always), nothing to be interpreted. Relying on the literal meaning is then not a matter of a normative choice and, in this sense, is amoral (Marmor does not use this word).
- 25 For the Polish discussion, see Grabowski 2015 in English.
34A different question is whether unreflective reliance on the literal meaning by judges is justifiable from the normative point of view. It may be argued that the application of law should always be deliberative, and a judge should avoid unreflective and uncritical acceptance of a literal meaning because such acceptance triggers the risk of wrong, unjust, or absurd decisions.25 Such a risk may be considered as a normative reason for adopting the maxim omnia sunt interpretanda, understood as a normative requirement of the following sort: “Even if a norm is linguistically clear, a judge should apply extralinguistic canons of interpretation to make sure that such a norm taken in this meaning is systematically, teleologically and morally right”. The positivistic answer to this objection is, roughly, that this is a normative concern, and therefore, it is not a matter of a descriptive theory of law.
35Marmor’s account is problematic for another reason. As I read Marmor, he assumes that understanding legal texts is the result of a cognitive process governed exclusively by linguistic intuitions grounded in linguistic competence and is therefore amoral. Such a view, although intuitively plausible, is not supported by the results of empirical research. Recent work in experimental jurisprudence has suggested that moral intuitions directly influence the linguistic understanding of a legal text. In particular, the results of experiments suggest that reliance on the literal meaning of legal norms by laymen is driven by their unreflective moral intuition relating to the requirement of the “publicity” of the law (Bystranowski, Hannikainen and Tobia 2024). Other empirical studies have suggested that one’s intuitive understanding is guided by their moral appraisal (Flanagan, de Almeida, Struchiner, and Hannikainen 2023) and that intuitive and tacit moral reasoning directly influences the understanding of legal texts and their application to the case under consideration (Kneer and Skoczeń 2023). Therefore, to confirm or refute Marmor’s conception of an “amoral” understanding of legal texts, further inquiry into the cognitive processes underlying the understanding of legal norms would be required. In particular, the question that should be answered regards what role (if any) tacit moral beliefs play in understanding legal norms.
36The empirical findings referred to above seem to “pose a challenge to positivist theories of law that advocate a strict conceptual separation between the domain of law and morality” (Flanagan, de Almeida, Struchiner and Hannikainen 2023). With respect to Marmor’s account, it may be argued that not only does legal interpretation engage in moral evaluations but so does the understanding of legal texts (either explicitly or tacitly). Therefore, the identification of legal content always depends on moral evaluations, which falsifies the separation thesis, even in its weak version.
37Let us consider in more detail the question of whether the results of the experiments referred to above are actually able to undermine the positivistic separation thesis. I believe the answer to this question depends on the answer to a more fundamental question, namely, how the reference to “morality” in the separation thesis is to be understood. Here, I have in mind the distinction between conventional (positive) morality and critical morality. This distinction is well known to legal philosophers, as it played a crucial role in the famous Hart–Devlin debate. Hart defined positive morality as the morality actually accepted and shared by a given social group, and critical morality as the general moral principles used in the criticism of actual social institutions, including positive morality (Hart 1963: 20). The subject-matter of ethics as a philosophical discipline is critical morality. For example, the dispute between consequentialism and deontology or (in practical ethics) the disputes relating to abortion or euthanasia pertain to critical morality. In ethical disputes, the question of which moral norms are accepted by members of a given group or society does not have much significance. Ethical disputes relate to the question what is right or wrong, not to the question what people consider to be right or wrong.
- 26 Not every natural law theorist would agree with this characterization of natural law theories. See, (...)
38Let us recall that the separation thesis is the main area of contention between legal positivism and the various natural law theories. For the latter, I assume that their common claim is that the law somehow necessarily depends on morality.26 No doubt, adherents of natural law theories have in mind critical morality and not conventional morality. In the claim that lex iniusta non est lex, the expression lex iniusta does not refer to the law that violates the morality contingently accepted by the members of a given society, but rather to the law that that violates the general moral principles of the “true” critical morality. Natural law theories are interested in what is genuinely morally right or wrong and not in what is considered right or wrong by members of society. Therefore, if legal positivists are to engage in a genuine dispute with natural law theorists, their separation thesis must also refer to critical morality. Otherwise, legal positivism and natural law theories would just be talking past one another.
39Even if the linguistic understanding of a legal text is guided by tacit moral intuitions, the separation thesis may remain true. The fact that people (both laymen and judges) tacitly rely on certain moral intuitions is simply a social fact, but this fact does not imply that these intuitions are correct in the sense of critical morality — they might be wrong in this regard.
40Two questions arise regarding such moral intuitions. The first is whether those intuitions are shared, in the sense that the overwhelming majority of laypeople have the same moral intuitions. If the answer is positive, we should ask what the basis of such shared intuitions is. Are they the hereditary outcome of the evolution of human minds? Or, have they been acquired through the process of socialization? If answer to the first question is positive, we are dealing with morality conceived of as a part of evolutionary human nature. If the latter answer is positive, then we are dealing with conventional morality. Nevertheless, in either case, we have to contend with descriptive facts, either biological or social. If in the determination of legal content people rely on innate (as opposed to conventionally shaped) moral intuitions, the understanding of the phrase “social facts” in the positivistic social thesis should be broadened to include the shared innate (and not only conventional) moral intuitions people possess.
41It may be the case that tacit moral intuitions guide the understanding of legal texts, but such intuitions are not uniform (different people have different intuitions), in which case such intuitions would not be the product of evolution or the effect of convention. This would not change the above conclusion. It would simply mean that the understanding of legal texts is guided by actual moral intuitions, the possession of which is a descriptive social (or psychological) fact.
42The fact that people rely on moral intuitions in the determination of legal content falsifies neither the separation thesis nor the social thesis. Refutation of the separation thesis (in its weak version) would require demonstrating that legal content cannot be determined without reference to critical morality or, in other words, without deliberation regarding what is morally right or wrong (as opposed to what people consider to be right or wrong). Experimental jurisprudence is unable to provide evidence for such a claim, as empirical research investigates what people actually believe and not what they should believe.
43The positivists may challenge the assumption that interpretation is always value-laden in various ways. I briefly discuss two of them in the following.
- 27 In continental legal theory, a similar view has been expressed, inter alia, by Duarte and Moniz Lop (...)
44(4.1) The positivists may assert that the traditional picture of the law is too narrow. Components of the law are not only statutory and customary rules (plus rules derived from precedential decisions in common law systems) but also canons of interpretation and rules of judicial reasoning (such as a contrario or a fortiori), provided that they are generally accepted in a given legal culture. For the sake of brevity, I refer to canons and rules of reasoning jointly as “canons” and I skip the problem of precedents as a source of law. Baude and Sachs claim that canons constitute “the law of interpretation”, which functions to determine the legal meaning of a statute in a legal system (Baude and Sachs 2017: 1580).27 They further claim that the cannons constitute unwritten law, which derives its validity from legal practice. The validity of the “law of interpretation” is ultimately a matter of social practice, on the same basis as the validity of other legal rules.
- 28 Baude and Sachs (2017: 1116) write: “We assume in this article something like Hartian positivism, p (...)
45The process of interpreting the law involves applying the law of interpretation. The law of interpretation satisfies the positivistic criteria of validity set out in the systemic rule of recognition.28 Thus, it is a part of the legal system. Of course, this does not per se solve the problem of evaluation, but Baude and Sachs claim that extracting legal content from a written document does not necessarily involve direct normative or evaluative judgments. Most of the normative choices at issue have already been made and are reflected in preexisting canons of interpretation, which have the status of legal rules (Baude and Sachs 2017: 1083). Therefore, the application of these canons are not based on the subjective normative choices of a particular interpreter but rather on the congruent practice of the legal community. Such a practice is a social fact.
46The difference between this account and the account defended by Marmor is clear. For Marmor, each interpretation involves an evaluation, but if the text is linguistically clear and no doubt arises from extralinguistic circumstances, no interpretation is required. For Baude and Sachs, canons of interpretation constitute a valid part of law, and therefore, as long as the determination of legal content is based on canons of interpretation, the legal content is ultimately a matter of social facts. Obviously, canons of interpretation are based on evaluative choices, but those choices are not made by the interpreter who works with canons that already exist as a part of the law. Baude and Sachs (2017: 1095) stress that “the legal system frequently chooses artificial rules of interpretation, and once chosen they’re the law, whether or not they reflect what a given text really meant”.
47Let us agree, for the sake of discussion, that the canons of interpretation are valid legal rules and that their validity is based on the conventions shared in a given legal culture. In this case, Baude and Sachs would be able to demonstrate that legal interpretation does not involve moral evaluations, provided that the application of the canons does not necessarily require a normative (moral or other) evaluation. The key problem lies in this proviso. If it is true, the weak separation thesis is true because, at least in some cases, determining legal content on the basis of the canons does not involve a moral evaluation. But is this proviso true?
48The problem of normative choices and evaluations in connection with the application of canons arises for two reasons. First, the canons sometimes refer to evaluative criteria for the determination of legal content (for example, “If a statutory clause has more than one linguistic meaning, choose the one which promotes the fair solution of the case”). I do not believe that such a reference to moral criteria by the canons triggers a problem for legal positivism. Exclusive positivists would likely argue that such a canon gives judges the power to legislate, whereas inclusive positivists would likely contend that in this way moral criteria are incorporated into the law.
49The second reason, namely the conflict of canons, is more serious. Recall Llewellyn’s claim that for each canon we may find a “duelling” one. He writes, “[T]here are two opposing canons on almost every point” (Llewellyn 1950: 401). If this is true, then the theory of the law of interpretation does not solve the problem of the separation thesis. Each interpreter, when applying a canon, explicitly or tacitly makes a normative choice, namely by selecting one of the duelling canons. Therefore, each determination of legal content involves an evaluation. Two scenarios are possible here: (i) The canons conflict in abstracto (for example, expresso uious est exlusio alteris and eiusdem generis), or (ii) the canons conflict only in concreto, with respect to a particular interpretative issue. For example, the canon of literal interpretation and the canon of purposive interpretation may conflict in respect to a particular case, but they may remain consistent in respect to other cases.
50On first consideration, the conflict between canons conceived of as valid legal norms does not seem to trigger a specific theoretical problem. It is commonplace that the law frequently contains conflicting norms. As far as legal rules are concerned, conflicts between them are usually resolved through application of so-called derogatory rules (for example, lex specialis derogate legi generali, or lex posterior derogate legi priori, or lex inferior not potest derogare legi superiori). Baude and Sachs would likely assert that such derogatory rules have the status of legal norms. Such rules, however, are useless with respect to conflicting canons of interpretation, as the criteria of specificity, hierarchy, and temporality are obviously not applicable to the canons. As far as legal principles in the Dworkinian sense are concerned, the conflicts between them are resolved through the procedure of balancing. But the canons of interpretation are not principles in the Dworkinian sense, as they are applicable in all-or-nothing fashion. A principle can be fulfilled to different degrees, whereas a canon of interpretation can be either complied with or not. There can be no balancing between conflicting canons.
- 29 I skip here the question of whether a particular interpretive doctrine can be incorporated into the (...)
51Therefore, the choice between conflicting canons cannot be made by applying derogatory rules or the balancing procedure. The choice is either made ad hoc (to achieve a particular outcome preferred by the interpreter) or is based on a normative doctrine of interpretation accepted (explicitly or tacitly) by the interpreter. Here, I have in mind such doctrines as formalism, textualism, intentionalism, and purposivism, among others, that are present in American legal culture (and their continental counterparts). Such doctrines may be conceived of as the sets of interpretative rules of the second level, the function of which is to solve conflicts between canons (interpretative rules of the first level). From the positivistic standpoint, the selection of a particular doctrine of interpretation is not a matter of convention: Such doctrines are not conventional.29 They are neither incorporated into the rule of recognition nor validated by the rule of recognition (Greenberg 2017). The acceptance of a doctrine of interpretation is rather a matter of political philosophy and is driven by the accepted values of political morality and their hierarchy. The doctrines govern the selection and use of the appropriate canons of interpretation. Therefore, even if canons of interpretation are a part of the legal system, their application for the purpose of determining legal content necessarily involves evaluations.
- 30 I skip Joseph Raz’s view here, who claims that as a matter of conceptual necessity, interpretation (...)
52(4.2) The claim that any choice of interpretative doctrine is a matter of political philosophy has been challenged by proponents of exclusive positivism. In the following, I briefly discuss only one prominent version of such an approach.30 Shapiro argues that legal reasoning (including interpretative reasoning) is necessarily amoral (Shapiro 2011: 266-7). This means that, as far as genuine legal reasoning is concerned, the interpreter does not invoke moral considerations. Law is a plan (or rather a system of plans): “[I]t would defeat the purpose of having plans if, in order to figure out whether a plan exists or what its content is, one had to resolve a question that the plan was designed to answer” (Shapiro 2011: 310). Insofar as law settles moral problems, the interpretation of law cannot be determined by morality. This claim is a consequence of the “general logic of plans” that Shapiro describes. Therefore, the choices between different interpretative methodologies made by an interpreter (between textual interpretation, intentional interpretation, purposive interpretation, and so on) cannot be based on moral arguments. The choices should be based instead on descriptive social facts, and the relevant social fact determining such a choice is the actual distribution of trust by the planner:
[T]he Planning Theory demands that the more trustworthy a person is judged to be, the more interpretative discretion he or she is accorded; conversely, the less trusted one is in other parts of legal life, the less discretion one is allowed (Shapiro 2011: 331).
53According to Shapiro, the identification of the actual distribution of trust is an empirical matter. Therefore, the choice of the appropriate interpretative methodology is a matter of social facts. As the application of the selected methodology determines the legal content, the legal content so determined is also indirectly a matter of social facts.
- 31 See, for example, Pino (2012) who argues inter alia that extracting the economy of trust from the p (...)
54Shapiro’s theory gives rise to many questions and objections. The main line of critique asserts that interpretation continuously faces the need to make normative choices and that such choices require moral evaluations.31 For Shapiro, the need for moral evaluation occurs only if the plan “runs out” and judges must engage in further social planning (Shapiro 2011: 276). Additionally, I would like to raise certain issues, which to my knowledge have not been extensively discussed in the literature to date. Shapiro claims that the legal content identified in the process of interpretation is a matter of social facts because the choice of interpretative methodology is based on the actual distribution of trust, and this distribution is a social fact. But interpretative methodologies are normative (because they state what interpreters should do). Therefore, there can be no logical entailment between the statements describing the actual distribution of trust and the normative methodology of interpretation. Thus, Shapiro falls victim to Hume’s Guillotine. To derive a normative methodology of interpretation from the description of the actual distribution of trust, a normative premise is required of a sort such as “One should follow the intention of the planner (lawmaker) with respect to the level of trust in law-applying agencies”. In my view, the adoption of such a premise would be based on a normative choice, and hence its justification would require recourse to moral reasoning. Therefore, Shapiro’s meta-interpretative theory is based on a hidden normative premise pertaining to political philosophy. Specifically, Shapiro’s theory must assume that the courts are obliged to follow the intentions of the lawmaker, at least with respect to the scope of discretion granted by the lawmaker. Several further normative problems arise in connection with this premise. For example, the question should be asked: Which lawmaker is meant here — historical, present, or rational? In the continental tradition, the answer to this question gives rise to the fundamental divide between so-called static and dynamic methodologies of interpretation.
- 32 Shapiro claims that the selection of the interpretative methodology by the meta-interpreter require (...)
55More importantly, Shapiro claims that the level of trust determines (through justifying the appropriate interpretative methodology) the scope of interpretative discretion. If the level of trust is high, the interpreter should select a methodology that affords them a broad scope of discretion. If trust is low, a methodology should be selected that eliminates discretion and forces the interpreter to strictly follow the wording. Let us assume that the actual level of trust is low. Therefore, a low-discretion methodology is required, but which one? There are several methodologies that restrict the discretion of interpreters in different ways — for example, textualism and intentionalism. Both methodologies come in several variants. The same issue of there being a variety of possible methodologies applies if the level of trust is high, as more than one existing methodology provides for a high level of discretion. Shapiro openly admits that certain types of textualism “confer roughly the same degree of discretion on interpreters as some purposive ones”. Therefore, he concedes that the principles he proposes “cannot inform the choice between such methodologies” (Shapiro 2011: 358).32
56One more issue arises in connection with Shapiro’s account. According to him, the planner is unable to anticipate all future cases. Therefore, when an interpreter is confronted with a case that has not been anticipated and its solution is not provided by the plan (when the law “runs out”), the interpretative methodology that reflects the actual distribution of trust does not allow for the identification of the legal content because such content does not actually exist. In such a case, the decision must be based on extra-legal (including moral) reasons (Shapiro 2011: 276). The interpreters must reach “outside the law”, which means that they act as a legislator.
57This suggests that we have to address two possible scenarios: either a solution to a given case is anticipated by the plan, or it is not. In the former scenario, the role of the interpreter is merely to discover a solution through application of the appropriate interpretative methodology. For Shapiro, interpretative methodologies provide for a wider or narrower scope of discretion. But how is that possible if discretion necessarily involves a choice between alternative solutions to an interpretative problem? If a solution is anticipated by the plan and must only be “discovered” by the interpreter, there is no room for discretion on the part of the interpreter because only one proper solution exists. The very concept of discretion assumes that more than one solution is possible. If there is no place for discretion, there is also no place for selecting an interpretative methodology based on the level of trust. To repeat: If the solution is provided by the plan, the role of the interpreter is simply to discover the solution. If the interpreter adopts a solution other than the one anticipated by the plan, the decision is wrong. In this scenario, there is no room for discretion. In the latter scenario (no solution is provided), the law-applying agency acts as a planner and not as an interpreter; therefore, interpretative methodologies are irrelevant.
58I think that the discrimination of situations where a judge acts as a plan applier versus as a plan designer (lawmaker) is really a matter of degree. If the law “runs out”, a judge usually does not refer exclusively to morality (or other external sources) to find a solution, but tries to find the solution through reference to constitutional principles, general principles of law, precedents, analogies, comparative research, and the like (Pino 2012: 196). Decisions in such cases are based on various legal reasons and not solely on moral or economic reasons. As Poscher contends:
[A]djudication remains a specifically legal enterprise even in cases of legal indeterminacy, even in hard cases for which the law holds no predetermined answer […] The creation of law in the process of adjudication does not become mere politics, economics or moral reasoning or legislation (Poscher 2017: 316).
59Hart stressed that
[the] law creating powers which I ascribe to the judges to regulate cases left partly unregulated by the law are different from those of a legislature: not only are the judges powers subject to many constraints narrowing his choice from which a legislature may be quite free[.] (Hart 2012: 273).
60The brief overview presented above demonstrates that legal positivism has a problem with developing a theory (or meta-theory) of legal interpretation fully consistent with its main tenets — in particular, with the separation thesis (even in its weak version).
61This problem does not arise for the deflationary position (proposed inter alia by Spaak but presumably tacitly or openly assumed by many positivists). According to this position, legal positivism should be perceived solely as a theory of legal validity. Therefore, the separation thesis (as a consequence of the social thesis) relates solely to the determination of validity and not to the determination of legal content.
62Marmor claims that legal positivists cannot accept the conjunction of the two claims that (i) law necessarily requires interpretation, and that (ii) interpretation necessarily involves recourse to moral evaluations. These two claims taken together would imply that the content of law can never be identified without recourse to moral evaluations, which, per Marmor, contradicts the separation thesis.
63In their defence of the separation thesis as applied to the determination of the content of law, legal positivists claim that either (i) the identification of legal content never involves moral evaluations, or (ii) the identification of legal content only contingently involves moral evaluations. Claim (i) is defended by Shapiro and (with some reservations) by Baude and Sachs, and claim (ii) is defended by Marmor, who agrees that although interpretation necessarily involves evaluations, not all laws require interpretation — in many (or even most) cases, an unreflective understanding is sufficient. Both claims trigger theoretical concerns — or at least I so argue.
- 33 On the distinction of theoretical and methodological positivism cf. Chiassoni (2008: 254), with ref (...)
64The conclusion that no positivistic theory of interpretation is conceivable is unsurprising, if one has in mind a normative theory of interpretation. Legal positivism aims to remain a descriptive theory, and it is therefore not positivism’s role to provide recommendations for how the law should be interpreted or how (pace Shapiro) the proper interpretative methodology should be selected. But this conclusion does not mean legal positivists should abandon their interest in legal interpretation. Interpretation is an important part of legal practice. There are no obstacles to developing descriptive theories of interpretation that would preserve the methodological principle of neutrality. Here, let us recall the distinction between theoretical and methodological positivism.33 Descriptive theories of interpretation are positivistic in the methodological sense. They do not have to assume theoretical positivism, in the sense that they may be true, notwithstanding whether legal positivism qua legal theory (theoretical positivism) is true.
- 34 Space does not allow me to elaborate on the problem of the universality of legal theory. See Raz 20 (...)
- 35 For the distinction between the external and internal point of view, see Hart 2012: 89.
65On first look, the view that legal positivism qua legal theory is confined to the matter of the validity of law and has nothing to say about interpretation seems to trigger a sort of internal inconsistency. On one hand, this view implies that legal positivists may give answers to the question of which norms are valid in a legal system, but on the other, they have no tools to answer the questions relating to legal content. I do not think that this inconsistency is genuine. Legal positivism qua legal theory defines itself as a universal theory of law and not as a theory of any particular legal system.34 As a general theory, legal positivism looks at the law from an external point of view, either radical or moderate.35 Legal positivists qua legal theoreticians make either external statements relating to any law (or to some sort of legal order — for example, contemporary municipal legal systems) or detached statements relating to a particular legal system (as opposed to committed). It is not the role of legal positivism qua general legal theory to ascertain the validity of any norms in any particular system; such ascertainment would require making an internal committed statement. Of course, legal positivists are free to make internal committed statements confirming or denying the validity of a norm in any particular system. Such statements would not, however, be part of the positivistic theory. Legal positivists, on the same basis, are free to make internal interpretative statements. Therefore, there is no inconsistency between the alleged capability of legal positivism to answer validity questions and its incapability to answer interpretative questions.
- 36 For a more detailed discussion, see Toh 2008.
- 37 This is the role Dworkin ascribed to legal philosophy. See, Dworkin 1986: 90.
66Another way of addressing this issue is to distinguish first-order and second-order questions (Toh 2021: 562). First-order questions concern whether there is a law addressing some particular question and its content. Second-order questions concern the nature of law and/or what the necessary features of law are. The role of legal positivism is solely to answer second-order questions. Hart believed that first-order statements are internal statements (statements made from the internal point of view) and, therefore, are normative in the sense that considerations regarding their truth (if they are truth-apt) or correctness must include some normative components.36 Legal positivism, shaped by the principle of neutrality, suggests that legal positivists are not players in the legal game but are rather observers of the game. Therefore, the positivistic legal theory is not “a silent prologue to any decision of law”.37
—Acknowledgment.— The research was founded by the National Science Centre, Poland (Harmonia Grant No. 2018/30/M/HS/00254). The author would like to thank two anonymous reviewers for their comments. Needless to say, responsibility for the views expressed herein, as well as for any errors, rests solely with the author.